The purpose is to determine the response rate and associated toxicity of combination tiazofurin and allopurinol in patients with refractory ovarian carcinoma. Epithelial ovarian carcinoma (EOC) will strike 24,000 women and cause 13,600 deaths in 1994. Despite aggressive cytoreductive surgery and available cytotoxic agents such as platinum and paclitaxel, EOC is the first leading cause of death from cancer in women and the leading cause of death from gynecologic neoplasms. It is imperative that new agents with sound biochemical rationales be tested. The trial will include patients with EOC who have failed platinum and paclitaxel and have elevated levels of inosine 5' monophosphate dehydrogenase (IMPDH) activity in their malignant ascites. On day 1-2 4400 mg/m2 of tiazofurin will be infused over one hour. If the IMPDH activity after two days is higher than 10% of baseline in the EOC cells extracted from ascites, the patient will be removed from the study. If the IMPDH activity is less than 10% of the baseline, patients will continue tiazofurin infusion on day 3-10 at a reduced dose of 2200 mg/m2. During the tiazofurin treatment the patient will receive 100 mg PO Q 4 hour of allopurinol to attain plasma hypoxanthine levels of 40-80 micromolar. If the hypoxanthine level is less than 40 micromolar then the allopurinol dose will be increased to 100 mg PO Q 3 hour. After completion of a 10 day course the patient will remain off drug for 21 days such that any one drug cycle will be 31 days. During the treatment cycle clinical chemistries and hematologic profiles will be monitored; if there is unacceptable toxicity or disease progression the tiazofurin plus allopurinol therapy will be stopped. The study will feature a two stage accrual design. If no responses are noted in the first 14 evaluable patients the study will be closed. If one or more responses are seen in the first 14 evaluable patients as many as 25 more evaluable patients will be accrued in an attempt to establish more exactly the response rate.